306 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS with this technique can be more readily obtained by observation of scale patterns. Serology and sexing Whilst the identification of group substances in blood (Culliford (23)) has now become a routine technique in forensic laboratories no such advances have been made with hair. Japanese workers have claimed (Yada, Mori and Okane (24) Yada, Ishimoto and Okane (25)) to be able to group single human hairs but experiments in British laboratories have been unable to confirm these claims. Work in this laboratory has not yielded a successful technique, mainly due to the false reactions with A and O anti-sera. It is still not clear where the group substances are located. Much work has been carried out on the sexing of human hairs (Phillips (26) Brinkman and Jobst (27)). Staining techniques were used to identify Y-chromosomes and Barr bodies and the results were very encouraging. The description of human hair colour Introduction It is a common observation that the appearance of human hair (a description which includes colour, texture and many other subjective factors) is a particularly characteristic feature of an individual. The colour of human hair is one of its most distinctive features but its measurement needs to be put on a less subjective basis. The purpose of this experiment was to develop a hair chart suitable for the study of the variation of head hair colours across the population. Samples of hair from 100 men and 100 women were collected. Experimental Hair from each individual was gathered into a small tuft of approxi- mately 50 hairs and mounted upon a small piece of neutral grey card. The colour of each tuft was compared with a Munsell hair colour chart (13) which contained matt painted chips arranged according to their hue, value and chroma. However, the results of this experiment demonstrated the inability of observers to match human hair (a highly textured material) to the matt surface chips used in the Munsell system. It is therefore essential that the colour of human hair is compared with coloured material of a similar surface texture. This is because the specular reflections from the

HUMAN HEAD HAIR AS FORENSIC EVIDENCE 307 surface of the hair constitute the large part of the colour sensation. The other major limitation in hair colour comparisons is that human hair dis- plays a very limited range of hues (581-606 nm), i.e. human hair differs mainly in the lightness or darkness of the colour, not in the wavelength of the reflected light. The 200 hair tufts were therefore compared with commer- cially available dyed nylon tufts (Wella (28)). Only 29 colour tufts were used to describe the colour of the hair samples, but on extending the colour com- parisons to single fibres a hair colour chart containing 29 colours merely led to ambiguity in assigning the hair colour. This is because the colour of single fibres from a given individual varies along the length of the fibre and across the head of that individual. Successive experiments indicated that nine hair colours (from black to white) provided sufficient colours to describe the population with minimum ambiguity in assigning the particular colour. It is neither necessary nor desirable to use more than nine colours to describe hair because of the range of hair colours displayed by a given individual. Some of the individuals examined displayed a range of colours from group 3-6 and 2-5 (Table I). Hence it is important to have sufficient controls to establish the range of hair colour displayed by a given individual. A single fibre was chosen from the hair of each of 200 individuals at random and compared with the hair chart containing nine colours. A hair was described as grey/white if the original tufts prepared contained too many white hairs to preclude colour comparisons. The results of this survey are shown in Table I. Table I. Distribution of hair colour in population Very dark Dark White Black brown brown Mid-brown Blonde •grey Colour group 1 2 3 4 5 6 7 8 9 Female % population 0 28 20 9 14 7 6 4 12 Male % population 0 43 16 12 9 1 0 0 19 Table I shows that a large proportion of the colours of the male and female population lie in groups 2, 3, 4, 5, 9. The remaining groups 1, 6, 7, 8 are sufficiently rare to provide useful discrimination. It can also be seen that no black hairs or genuine red hairs were found in this survey. So-called blonde hair (groups 6, 7, 8) were sufficiently rare in both male and female